Identification and consequences of distinct Loe-Silness gingival index examiner styles for the clinical assessment of gingivitis

Background: In clinical studies, gingivitis is most frequently assessed by the Loe-Silness gingival index (GI). The objective of this work was to deve lop an understanding of how clinicians experienced with GI differ with resp ect to how they apply GI and to assess the impact of different examination styles on statistical outcomes and magnitude of treatment differences. Methods: A method was developed to mathematically relate the average GI sco re and degree of bleeding observed for a subject. Graphical analyses were u sed to profile examiner styles with respect to using the GI index. A prospe ctive single-center, examiner-blind study comparing the effects of a stagge red prophylaxis on gingivitis was then conducted, where a difference in gin givitis was created between two balanced groups by providing subjects a pro phylaxis at two staggered time points. Subjects were assigned to one of two cohorts; within each cohort, group 1 subjects received a dental prophylaxi s following the baseline examination and group 2 subjects received a dental prophylaxis 8 weeks later. Five to 7 days after the group 2 prophylaxis, a ll subjects were examined for GI. Twelve experienced clinicians participate d. Results: Retrospective analyses indicated the presence of distinct examiner styles which are based on the frequency that a given GI score (0, 1, 2, or 3) is measured by a clinician. In the prospective study, all 12 examiners observed statistically significant differences between the prophylaxis trea tment groups at the final visit for both mean number of bleeding sites and mean GI; the magnitude ranged from 21.5% to 84.6% for mean number of bleedi ng sites and 9.4% to 39.2% for mean GI. There were 4 distinct styles employ ed by these experienced clinicians. Conclusions: Varying examiner styles impact the structure of resulting data . Importantly, the implementation of arbitrary thresholds (e.g., 20%) regar ding percent treatment differences between groups as a guideline for judgin g the clinical significance is scientifically unsupported. A more scientifi c criterion in the field of gingivitis clinical testing would be the indepe ndent demonstration of statistical superiority compared to a negative contr ol and/or a demonstration of similar or superior efficacy to clinically pro ven positive controls. In addition, interexaminer calibration is a mechanis m that can be utilized to minimize the impact of different examiner styles in clinical settings involving more than one examiner.